All Categories - SHDB

Summary
At the Social Hotspots Database (SHDB) project we were recently made aware of the Orienting project report on databases for social LCA. We welcome and concur with the report’s emphasis on the need for transparency in data for LCA including for social LCA. We also therefore agree with the emphasis on transparency that comes in a recent comment posted by Andreas Ciroth of GreenDelta, the seller of the PSILCA database which was reviewed in the Orienting project’s report along with the SHDB.

However, we regret that in his comments Ciroth made a series of false claims about the Social Hotspots Database (SHDB), especially in relation to this key topic of transparency.  For this reason, the research and data team at SHDB has written a careful and detailed rebuttal.   Please read our full post, but here are some key points:
The SHDB has always been 100% transparent about every source and every step in the methodology and database, and it continues to be so. The SHDB is precisely an implementation of the Life Cycle Attribute Assessment approach to Social LCA, which was introduced in transparent peer-reviewed journal articles by the creators of the SHDB.

The SHDB’s full detailed documentation, made available to every user of the database, is 150 pages in length (not 30). Its 100% transparency has repeatedly been cited by users as a reason to use the SHDB instead of less transparent data sources such as Maplecroft. This 100% transparent document cites (with links) every one of the more than 200 data sources. Every characterization model which estimates risk based on primary data sources is explicitly detailed in the documentation.  The full 150-page documentation is also made available to other researchers upon request.  The 100% transparency enables every researcher to replicate the risk evaluations in our transparently published characterization equations, AND to use alternative risk assessments if/as they see fit.
The SHDB also provides more than risk-assessed worker-hours, contrary to another false claim by Ciroth. The SHDB also includes skilled and un-skilled worker hours, as well as a full set of socio-economic variables including wage payments to 5 different occupational classes, value of capital inputs, value of land inputs, value of other natural resource inputs, and total value added.

Users of OpenLCA and SimaPro can easily introduce edited risk level evaluations into copies of the SHDB unit processes which they can freely create in these software tools, contrary to Ciroth’s false claim that users “cannot change risk levels.”

What Ciroth attempts to show as “evidence” of SHDB non-transparency is screenshots of the OpenLCA software he has a financial stake in, which has been specifically engineered to match the structure of the PSILCA database he has a financial stake in.  To the casual observer of his screenshots, his PSILCA may appear more transparent than the SHDB when viewed through his software.  However, an engineered simulation of additional transparency does not negate the reality of 100% transparency already provided by the SHDB.

Finally, Ciroth speculates at length about conflicts of interest on the part of one of the Orienting authors, Mark Goedkoop. However, both the SHDB and PSILCA are available for use in both OpenLCA (in which Ciroth has a financial stake) and SimaPro (from the company PRe that Mark Goedkoop founded but is no longer CEO of). Ciroth confirms this himself with his own screenshots. We find it straightforward to conclude that if Mr. Goedkoop would still have any financial incentive to promote the SHDB (of which we are not aware), then he would have equivalent financial incentive to promote PSILCA. So we find no reason to suspect bias or conflict of interest. The authors of the report have simply documented the equivalence in the transparency of the two databases.

In addition, the SHDB has specific strengths in its risk and trade modelling methodology that differentiate it. These strengths include 1) using a mix of qualitative and quantitative indicators to bolster the robustness of the risk assessment, supported by a social science-based risk assessment rubrics, 2) the integrated GTAP global Input-Output model have the greatest granularity at the summarized level (65 sectors consistently being represented across 160 countries).

More Detailed Response
The research and data team at NewEarthB, which has created and continually updates and expands the Social Hotspots Database, is writing to provide a detailed and factual rebuttal to the series of unfortunately false claims about the SHDB recently made by Andreas Ciroth.

Claim 1: Ciroth claims that the SHDB “contains only risk-assessed worker hour flows.”
Fact 1: The SHDB provides more than risk-assessed worker-hours including a categorization between the skilled and unskilled worker hours. It also includes a suite of socio-economic variables including:
Capital (USD/USD) Land (USD/USD) Natural resources (USD/USD) Wages to clerks (USD/USD) Wages to low skill labor (USD/USD) Wages to service labor (USD/USD) Wages to tech labor (USD/USD) Wages to office/manager labor (USD/USD) Value added (USD/USD) Work hours
Claim 2: Ciroth claims that the SHDB is “not sufficiently transparent.”  He further claims that our documentation is only 30 pages.

Fact 2a: Although Ciroth fails to define what “sufficient” transparency would be, this is no matter because the SHDB is, and always has been since its very first version, 100% transparent.  First, it (and PSILCA too) are both implementations of the Life Cycle Attribute Assessment approach to Social LCA, which was introduced in transparent peer-reviewed journal articles by the creators of the SHDB Norris 2006, Andrews et al. 2009, Benoit Norris et al. 2012.)

Fact 2b: Our full detailed documentation, which is 150pages in length (not 30), cites and gives links to each of the more than 200 data sources  and every characterization model which estimates risk based on  primary data sources, is explicitly detailed in our 150page long documentation which is made available to every user of the SHDB, and to other researchers upon request.  This 100% transparency enables every researcher to replicate our risk assessments, AND to use alternative risk assessments if/as they see fit (see below).
The trade flow model in the SHDB is also transparently based on the extensively peer-reviewed GTAP model (https://www.gtap.agecon.purdue.edu/about/project.asp ).

The SHDB’s 100% transparency has undoubtedly been an enabler of more than one follow-on project that has offered competing data source in the market, but we at SHDB believed from the start that risking such replication was something we were compelled to do in order to pioneeringly promote and advance the field and practice of LCAA, of Social LCA, and of wider efforts to improve working conditions in supply chains. We still believe this and therefore continue to provide it.

Claim 3: Ciroth claims that “users cannot change the risk levels of indicators.”
Fact 3: This too is patently false, as Ciroth is surely aware. First, as noted in Fact 2b, all of the underlying primary data are transparently documented, as are the risk characterization models.  Users have the ability to change the risk characterizations of the SHDB based on their research and data and to propose and use their own –, And as Ciroth surely knows, users of OpenLCA and SimaPro can easily introduce edited risk level evaluations into copies of the unit processes which they can freely create in these software tools.

This 100% transparency and 100% flexibility/adaptability has been used in countless consulting projects and many research projects around the world; see for example the work of Blackstone et al., 2021.

What Ciroth attempts to show as “evidence” of SHDB non-transparency is screenshots of a software he controls and has a financial stake in, which has been specifically engineered to match the structure of the PSILCA database he has a financial stake in.  To the casual observer, the database he has a commercial interest in looks more transparent than the SHDB when viewed through the software interface he has a commercial interest in.  However, an engineered simulation of additional transparency does not negate the reality of 100% transparency already provided by the SHDB.

Claim 4: Ciroth claims that SHDB lacks a data quality framework.
Fact 4: The SHDB uses the same data quality framework as of PSILCA. It was introduced in 2022 and can be found in the SHDB documentation.

Claim 5: Ciroth speculates at length about conflicts of interest on the part of one of the Orienting authors, Mark Goedkoop.

Fact 6: Both the SHDB and PSILCA are available for use in both OpenLCA (in which Ciroth has a financial stake) and SimaPro (from the company PRe that Mark Goedkoop founded but is no longer CEO of). Ciroth confirms this himself with his own screenshots. We find it straightforward to conclude that if Mr. Goedkoop would still have any financial incentive to promote the SHDB (of which we are not aware), then he would have equivalent financial incentive to promote PSILCA. Therefore we find no reason to suspect bias or conflict of interest. The authors of the report simply documented the equivalence in the transparency of the two databases.

Citations
Norris, GA, 2006: “Social Impacts in Product Life Cycles: Towards Life Cycle Attribute Assessment”. Int. JLCA, 11(1):97-104.
Andrews Evan, Lesage Pascal, Benoît Catherine, Parent Julie, Norris Gregory, Revéret Jean-Pierre. LifeCycle Attribute Assessment: Case Study of Quebec Greenhouse Tomatoes. 2009. Journal of IndustrialEcology, Volume 13, Issue 4.
Benoit-Norris, C, DA Cavan, and GA Norris, 2012. Identifying Social Impacts in Product Supply Chains: Overview and Application of the Social Hotspot Database. Sustainability 4: 1946-1965.
Blackstone, NT, C Benoit Norris, T Robbins, B Jackson, JL Decker Sparks, 2021. Risk of forced labour embedded in the US fruit and vegetable supply. Nature Food. https://doi.org/10.1038/s43016-021-00339-0